Most PM programs get built once, usually from the equipment manual, and then they run on the calendar for years. Monthly, quarterly, annual. The asset doesn't know what month it is. It knows how hard it's run, how dirty the area is, how old it is, and what was done to it last time.
That gap is where PM programs quietly go wrong. Not because anyone is lazy, but because a schedule that was reasonable on day one never gets tested against what the equipment actually does.
What a calendar PM assumes
Every time-based PM rests on three assumptions:
• The interval is about right.
• The task would catch the failure it is meant to prevent.
• Someone writes down what they found, so you can tell the first two are true.
Most sites never check any of the three. The PM gets done, the compliance number looks fine, and nobody asks whether the work changed anything about how the machine fails.
I've picked up PMs that were copied from the manual word for word and never touched again. Nobody was hiding anything. The task list was just never compared to the failures the plant was actually having.
The audit: three questions per asset class
You don't need a consultant or new software. You need the last year of closed PMs and the corrective work orders for the same assets. Pick one asset class at a time, like gearboxes, conveyors or air compressors.
1. What did this PM find?
Count the closed PMs where the tech found something and the ones where nothing was found. A PM that finds nothing every single time is one of three things: the interval is too short, the task is checking the wrong thing, or it was closed without anyone really looking. You can't tell which unless the closeout says what was checked and what the condition was.
2. What failed between PMs?
Line up the corrective work orders next to the PM tasks. A failure with no matching PM task is a gap in the program. A failure on an asset that just had its PM means the task isn't catching what it should. Either way, that is information the calendar will never give you on its own.
3. What has changed since the PM was written?
Equipment gets rebuilt, moved to a different duty, run more shifts, or relocated to a dirtier area. The PM written for the original conditions keeps running the same way. Ask when the interval and task list were last reviewed against how the asset is actually used. If the answer is "at install," that is your starting point.
What to do with the answers
Keep the changes small and one asset class at a time:
• Where PMs find nothing for a long time and nothing fails between them, the interval can probably stretch. Try it on one class first and watch it.
• Where failures keep showing up between PMs, shorten the interval or add the task that would have caught it.
• Where the failure gives warning, swap a calendar task for a condition check. Vibration, temperature, oil condition or a simple inspection point can tell you more than a fixed date.
• Where nobody can say what a task catches, question it. If no one can name the failure it prevents, it may just be filling the schedule.
The point isn't to do less PM work. It's to put the hours where they find something.
Where this usually breaks
The whole audit depends on what the tech wrote at closeout. If most closed PMs say "done" or nothing at all, all you can measure is compliance, and compliance is the number that looked fine to begin with.
If your closeouts are thin, fix that first. Count a week of them. Read a few back to the techs who wrote them. A note that someone reads within the week is the thing that makes the next note better. The audit gets easier every month after that, because the history starts to mean something.
A small way to start
Take the ten assets that cost you the most downtime last year. Pull their PM tasks and their corrective history side by side, on one page. Look for two things: failures no PM task covers, and PMs that have found nothing in a year. That is one afternoon of work, and it will show you more about your program than another round of compliance reports.
A PM program is a set of guesses about how equipment fails. The audit is how you find out which guesses were right.
